how to get layer-to-layer output in DL workbench

[ShaLu]

Hi, 

I noticed (in this post https://github.com/openvinotoolkit/openvino/issues/1551) that layer-to-layer outputs can be examined using OpenVINO DL workbench.  It is very useful for checking the workflow of the model. Could anyone know how to get the layer-to-layer outputs in the DL workbench? Thanks!

Regards,

Sha

[Wan_Intel]

Hi ShaLu,

We are investigating this issue, and we will update you at the earliest.

Thank you.

Regards,

Wan

[Wan_Intel]

Hi ShaLu,

Thanks for your patience.

For your information, after you have created your project in OpenVINO™ Deep Learning Workbench, you may visualize the layer with the following steps:

1.   Click “Open”

2.   Click “your model”

3.   Click “Visualize Model”

 

Regards,

Wan

[ShaLu]

Hi Wan,

Thanks for your reply.

It helps to see the structure of my model. However, what I am looking for is to see the outputs of each layer when making an inference. Details can be found in the next reply I posted.

Regards,

Sha

[ShaLu]

Hi Wan,

Thanks for your reply. 

Following the steps you posted, I can visualize my model as the following.

However, I would like to see the outputs of each layer when making an inference, like the outputs shown in the post https://github.com/openvinotoolkit/openvino/issues/1551, which could look like this.

I have explored the DL workbench but couldn't find where to see the step-by-step outputs. I am appreciated that if you could help me with this.

Regards,

Sha

[Wan_Intel]

Hi ShaLu,

For your information, referring to the GitHub question, the customer created a custom Python script to view the output results. The output result cannot be viewed in OpenVINO™ Deep Learning Workbench. You may refer to "Unit-test code" in this question for more information.

Regards,

Wan

[Wan_Intel]

Hi ShaLu,

Thanks for your question.

This thread will no longer be monitored since we have provided information. 

If you need any additional information from Intel, please submit a new question.

Best regards,

Wan